Sequence Image Match Based on Salient Point Invariants Moments
Junchai Gao, Zhiyong Lei, Zemin Wang, Hanshan Li · International Journal of Digital Content Technology and its Applications · 2012
Abstract In order to address estimation of mobile robot navigation, image sensors need acquire environmental information. This paper proposes sequence image match based on salient point invariants moments. It identifies salient point with SIFT scale-invariant algorithm, and the salient points have scale-invariant properties; it computes invariants moments of salient point neighborhood, search for matching points using KD-tree algorithm; the mismatch points are removed with RANSAC algorithm, and then it calculates image transformation matrix. Experiments show that, the proposed image match algorithm is not sensitive to scale and light, which is an effective image match method.